Classification Approach for COVID-19 Gene Based on Harris Hawks Optimization

A number of COVID-19 outbreak classification and prediction methods have been proposed and are being applied around the globe to make the right decision and to enforce proper control measures. Among these methods, simple statistical and epidemiological methods have received much attention whereas, for the long term prediction, the standard methods do not perform well due to the lack of essential data and high-level of uncertainty. Thus, the essential robustness and generalization abilities of these methods need to be improved. Therefore, this work proposed a new hybrid Harris Hawks Optimization (HHO) combined with the Support Vector Machine (SVM) method called HHO-SVM. The HHO-SVM is applied on a big Gene Expression Cancer (RNA-Seq) dataset which comprises more than 20531 features to identify the critical Gene that causes the COVID-19. The experimental results revealed that HHO-SVM outperformed than Grey Wolf Optimizer (GWO), Particle Swarm Optimization (PSO), Moth Flame Optimization (MFO), Slap Swarm Algorithm (SSA) and Genetic Algorithms (GA’s). We further investigate that the most critical Gene is Tmprss2 which causes Prostate Cancer is the same Gene that causes COVID-19 through the ACE2 receptor.